Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Former Tesla Firmware Engineer Discusses the System

twitter.com

Post reply on HN
Print view View original

341–350 of 589 posts

Re: Former Tesla Firmware Engineer Discusses the System

#341

Post by scrollaway »

debt wrote: post #226
ryandrake wrote: post #194

Earlier quoted context omitted.

I heard the same thing about the burnout and turnover. Do companies really think they are saving money by paying peanuts, grinding people down to burnout, and then constantly having to rehire/retrain new people as the old ones leave? Meanwhile the code is a mess because nobody has been there longer than a year and there is no architecture or design continuity. Just frantic patches over other frantic patches by burnt…

“Do companies really think they are saving money“ Well they landed a rocket on a barge in the ocean so something about that model must be working right

Looking at what's been done isn't a good way of determining whether your process is efficient at doing said thing. You need to be able to compare it to something else.

To put it another way: If your method of writing novels is to hire an infinite amount of monkeys and put them to work on typewriters, you can't say "Something about this model must be working right, I came out of it with the complete works of Shakespeare!"

They landed a rocket on a barge in the ocean. Maybe with a better process, they could have done that two years faster, for 1/100th the cost, with no burnout. You don't know, and you can't say the model works right just because there's something to show for it.

All you know is that the process is able to eventually land a rocket on a barge. It doesn't tell you whether it's good at it.

Re: Former Tesla Firmware Engineer Discusses the System

#342

Post by Cyph0n »

coldtea wrote: post #146
mlaretallack wrote: post #99

Earlier quoted context omitted.

Agree completely. Hard real time is possible with Linux, we use it for sub millisecond control of Traffic Lights. The only issue we ever hit is proving that the code running is the stuff we expected to run.

> we use it for sub millisecond control of Traffic Lights Why would anyone need "sub millisecond control of Traffic Lights"? Traffic lights are mission critical systems, of course, but even millisecond precision should be more than enough, and possibly even 0.5x-1 second precision...

The real question is: why do you need an entire OS kernel to control traffic lights?

Re: Former Tesla Firmware Engineer Discusses the System

#343

Post by analognoise »

BugsJustFindMe wrote: post #322
gascan wrote: post #256

Earlier quoted context omitted.

Is that like, you really don't need a fancy RTOS for this stuff 99.999% of the time, although sometimes you do? Or truly, despite the life safety element, there is never any need for RTOS.

In every application I can think of off the top of my head, but mostly in the ones that apply to Tesla, you truly don't, except when the law or a contract says otherwise. I'm sure there are exceptions for things that don't apply to Tesla (or for that matter SpaceX). Here's what's inside of every autonomous vehicle ever made: a message-passing subsystem, sensors, fusers, navigation, dynamic control, actuator device dr…

I think in terms of missiles, so everything needs an RTOS. Or bare metal.

Re: Former Tesla Firmware Engineer Discusses the System

#344

Post by dnbgfher »

BugsJustFindMe wrote: post #74
8_hours_ago wrote: post #52

Earlier quoted context omitted.

> the rockets fly with a heavily customized Linux install My jaw hit the floor the first time I heard this. Why Linux instead of an RTOS?? Apparently Tesla's autopilot also runs Linux, which seems like a huge accident waiting to happen (pun intended).

As a person who used to write navigation and control software for autonomous vehicles who occasionally gets downvoted for telling people that you really don't need a fancy RTOS for this stuff, you really don't need a fancy RTOS for this stuff. Linux is a very common platform for highly responsive robotic systems. I promise that their pid control isn't an Electron app.

Having done just enough robotics stuff to have used linux before (though not controlling anything even kind of large), I find this an odd notion. It's not that I think linux would go wrong for this sort of thing often, but when dealing with things like rockets or self-driving cars I'd think you want more assurance than "well we haven't had it be slow yet."

I haven't messed with an RTOS before but have done some fooling around with scheduling on microcontrollers and I can see why linux is tempting for ease and speed. But we're talking rockets and self-driving cars. These things are expensive as hell, can easily kill people, or both. It seems like the exact sort of place you'd want to take the time and effort to be sure.

Re: Former Tesla Firmware Engineer Discusses the System

#345

Post by T2_t2 »

Apocryphon wrote: post #272
watwut wrote: post #232

Earlier quoted context omitted.

Burning out people at large rate is not effective and persistent crunch is contraproductie. It is less of capitalism and more of wishfull thinking combined with wish to be seen as tough manager.

I'd say it's the pace in which modern capitalism operates at, particularly in tech. When the rate of production goes higher and higher due to automation and a global outsource-able workforce, the shareholders chase after higher and higher profits.

Nothing has changed, and ironically, this whole part of the question - is it better to burn people out and rehire - is a capitalist question, and answered with capitalist goals.

Re: Former Tesla Firmware Engineer Discusses the System

#346

Post by njloof »

Avshalom wrote: post #30
sz4kerto wrote: post #8

Why is this interesting or surprising? Of course almost all challengers have hacky tech under the hood because they don't have the resources. Winning from that position is possible not because of generally better tech but by delivering something that the incumbents don't. Remember the demo of the first iPhone (vs the huge expertise of Nokia). Or how Microsoft won the desktop starting from a single user, cooperative m…

because about twice a thread some one claims Tesla Has A Software Culture as a reason that GM or Toyota or VW won't ever be able to compete with them.

They didn’t say it was a good software culture.

Re: Former Tesla Firmware Engineer Discusses the System

#347

Post by beltex »

While were on the topic:

USENIX talk from VP of software at SpaceX (2016)

https://www.usenix.org/conference/lisa16/conference-program/...

(Speaker was concurrently interim VP of Autopilot at Tesla for a bit as well)

Few interesting bits:

- Speaker’s background is SRE at Google

- Engineering culture: Authority, autonomy, accountability. Blameless postmortems

- High Integrity C++ & MISRA coding standards

- The spacecraft has multiple onboard computers, all running Linux

- Tripe String Architecture, 3 redundant computers, whose results are cross checked with majority voting before being applied in real time, given radiation tolerant over radiation hardened hardware

Re: Former Tesla Firmware Engineer Discusses the System

#348

Post by jethro_tell »

gascan wrote: post #263
bradleybuda wrote: post #49

At DEF CON 23 there were two car-hacking talks that made an impression on me: one about hacking a Tesla [1] and one about hacking a Jeep [2]. FWIW, I came away thinking that the software architecture of the Tesla was light-years ahead of what Jeep (and presumably many other legacy manufacturers) was shipping: [1] https://www.youtube.com/watch?v=KX_0c9R4Fng [2] https://www.youtube.com/watch?v=OobLb1McxnI

Honestly I care a lot more about reliable function than security, at least as long as attackers can't get in wirelessly. Having both would be nice, of course.

>at least as long as attackers can't get in wirelessly

Do you not have bluetooth, satellite, or LTE in your car?

Re: Former Tesla Firmware Engineer Discusses the System

#349

Post by jacques_chester »

ummonk wrote: post #329
samstave wrote: post #243

Earlier quoted context omitted.

On a tangent, but still germane to the thread: Why arent their efforts to create "autonomous only" traffic managment scenarios, where people drive into a given, known area, and the area then takes control of managing the traffic and vehicles. Such that you relinquish control of the vehicle to that area's control system, with your destination stated and then your vehicle is managed accordingly. For example, a parking…

If it were that easy to setup tech-based infrastructure, we would have had positive train control implemented years ago.

Rio Tinto have spent more than a decade on autonomous freight trains.

On tracks they own.

With no other traffic than theirs.

And unlimited funding.

Re: Former Tesla Firmware Engineer Discusses the System

#350

Post by digi_owl »

ddevault wrote: post #11

I used to work at SpaceX on the team that did a piece of software called "WarpDrive". It was a massive monolithic ASP.NET application, with large swaths done with ASP.NET WebForms and a slow frontier of ASP.NET MVC gradually growing when I was working there. This application was responsible for practically everything that ran the factory: inventory, supply chain management, cost analysis, etc. Elon is a big Windows f…

> Thankfully the rockets fly with a heavily customized Linux install.

And this is why Torvalds takes kernel API stability seriously...

Post reply on HN